cloud storage as the infrastructure of cloud computing

7/31/2019 Cloud Storage as the Infrastructure of Cloud Computing

1/4

Cloud Storage as the Infrastructure of Cloud Computing

Jiyi WU1,2, Lingdi PING1, Xiaoping GE3,Ya Wang4, Jianqing FU1

1.School of Computer Science and Technology, Zhejiang University

2.Alibaba Business College, Key Lab of E-Business and Information Security,Hangzhou Normal University

3.Department of Electronics Information,Hangzhou Academy of Scientific Professional Technology4.Zhejiang Second Institute of Surveying and Mapping

Hangzhou, P.R.China

e-mail: [email protected]

AbstractAs an emerging technology and businessparadigm, Cloud Computing has taken commercialcomputing by storm. Cloud computing platforms provideeasy access to a companys high-performance computing andstorage infrastructure through web services. With cloudcomputing, the aim is to hide the complexity of ITinfrastructure management from its users. At the same time,cloud computing platforms provide massive scalability,99.999% reliability, high performance, and specifiableconfigurability. These capabilities are provided at relativelylow costs comparedto dedicated infrastructures.This article

gives a quick introduction to cloud storage. It covers the keytechnologies in Cloud Computing and Cloud Storage, severaldifferent types of clouds services, and describes theadvantages and challenges of Cloud Storage after theintroduction of the Cloud Storage reference model.

Keywords-Cloud Computing; Cloud Storage; architecture;

service; virtualization

I. INTRODUCTIONCloud computing portends a major change in how to

store information and run applications. Instead of runningprograms and data on an individual desktop computer,everything is hosted in the clouda nebulousassemblage of computers and servers accessed via the

Internet. Cloud computing lets you access all yourapplications and documents from anywhere in the world,freeing you from the confines of the desktop and makingit easier for group members in different locations tocollaborate.

Providers such as Amazon, Google, Salesforce, IBM,Microsoft, and Sun Microsystems have begun to establishnew data centers for hosting Cloud computingapplications in various locations around the world toprovide redundancy and ensure reliability in case of sitefailures. Since user requirements for cloud services arevaried, service providers have to ensure that they can beflexible in their service delivery while keeping the usersisolated from the underlying infrastructure. Recentadvances in microprocessor technology and software have

led to the increasing ability of commodity hardware to runapplications within Virtual Machines (VMs) efficiently.VMs allow both the isolation of applications from theunderlying hardware and other VMs, and thecustomization of the platform to suit the needs of the end-user. Providers can expose applications running withinVMs, or provide access to VMs themselves as a service(e.g. Amazon Elastic Compute Cloud) thereby allowingconsumers to install their own applications.Whileconvenient, the use of VMs gives rise to furtherchallenges such as the intelligent allocation of physical

resources for managing competing resource demands ofthe users.

One of the primary uses of cloud computing is for datastorage. With cloud storage, data is stored on multiplethird-party servers, rather than on the dedicated serversused in traditional networked data storage.When storingdata, the user sees a virtual serverthat is, it appears as ifthe data is stored in a particular place with a specificname. But that place doesnt exist in reality. Its just apseudonym used to reference virtual space carved out of

the cloud. In reality, the users data could be stored on anyone or more of the computers used to create the cloud.The actual storage location may even differ from day today or even minute to minute, as the cloud dynamicallymanages available storage space. But even though thelocation is virtual, the user sees a static location for hisdataand can actually manage his storage space as if itwere connected to his own PC.

Cloud storage has both financial and security-associated advantages. Financially, virtual resources in thecloud are typically cheaper than dedicated physicalresources connected to a personal computer or network.As for security, data stored in the cloud is secure fromaccidental erasure or hardware crashes, because it isduplicated across multiple physical machines; since

multiple copies of the data are kept continually, the cloudcontinues to function as normal even if one or moremachines go offline. If one machine crashes, the data isduplicated on other machines in the cloud.

II. KEY TECHNOLOGIESA. Cloud Computing

Cloud computing is about moving services,computation or datafor cost and business advantageoff-site to an internal or external, location-transparent,centralized facility or contractor. By making dataavailable in the cloud, it can be more easily andubiquitously accessed, often at much lower cost,increasing its value by enabling opportunities for

enhanced collaboration, integration, and analysis on ashared common platform.

Depending on the type of provided capability, thereare four scenarios where Clouds are used as showed inFig.1:

1)Infrastructure as a ServiceIPs manage a large set of computing resources, such

asstoring and processing capacity. Through virtualization,they are able to split, assign and dynamically resize thesere-sources to build ad-hoc systems as demanded bycustomers,the SPs. They deploy the software stacks that

2010 International Conference on Intelligent Computing and Cognitive Informatics

978-0-7695-4014-6/10 $26.00 2010 IEEE

DOI 10.1109/ICICCI.2010.119

380


2/4

run their ser-vices. This is the Infrastructure as a Service(IaaS) scenario.

2)Platform as a ServiceCloud systems can offer an additional abstraction

level:instead of supplying a virtualized infrastructure, theycanprovide the software platform where systems run on.Thesizing of the hardware resources demanded by the

execu-tion of the services is made in a transparentmanner. Thisis denoted as Platform as a Service (PaaS). Awell-known example is the Google Apps Engine.

Figure 1. Cloud Computing service types with examples

3) Storage as a ServiceCommonly known as Storage as a Service (StaaS),it

facilitates cloud applications to scale beyond their limitedservers. StaaS allows users to store their data at remotedisks and access them anytime from any place.Cloudstorage systems are expected to meet several rigorousrequirements for maintaining users data and information,

including high availability, reliability, performance,replication and data consistency;but because of theconflicting nature of these requirements, no one systemimplements all of them together.

4) Software as a ServiceFinally, there are services of potential interest to a

widevariety of users hosted in Cloud systems. This is analter-native to locally run applications. An example of thisisthe online alternatives of typical office applications suchasword processors. This scenario is called Software as aSer-vice (SaaS).

B. Cloud StorageCloud storage is amorphous today, with neither a

clearly defined set of capabilities nor any single

architecture. Choices abound, with many traditionalhosted or managed service providers (MSP) offeringblock or file storage, usually alongside traditional remoteaccess protocols or virtual or physical server hosting.Other solutions have emerged, typified by the Amazon S3service, that resembles flat databases designed to storelarge objects.

The Taneja Group defines cloud storage as a specificcategory within the larger field of storage in the clousolutions. Storage in the cloud encompasses traditionalhosted storage, including offerings accessed by FTP,

WebDAV, NFS/CIFS, or block protocols either remotelyor from within a hosted environment. Cloud storage is anevolution of this hosted storage technology that wrapsmore sophisticated APIs, namespaces, file or data locationvirtualization, and management tools, around storage. Fig2. shows the evolution of Cloud Storage based ontraditional network srorage and hosted storage.

There are hundreds of different cloud storage systems.Some have a very specific focus, such as storing Web e-mail messages or digital pictures. Others are available tostore all forms of digital data. Some cloud storage systemsare small operations, while others are so large that thephysical equipment can fill up an entire warehouse. Thefacilities that house cloud storage systems are called datacenters.

At its most basic level, a cloud storage system needsjust one data server connected to the Internet. A client(e.g., a computer user subscribing to a cloud storageservice) sends copies of files over the Internet to the dataserver, which then records the information. When theclient wishes to retrieve the information, he or sheaccesses the data server through a Web-based interface.

The server then either sends the files back to the client orallows the client to access and manipulate the files on theserver itself.

Figure 2.Evolution of Cloud Storage

Comedian George Carlin has a routine in which hetalks about how humans seem to spend their livesaccumulating stuff. Once they've gathered enough stuff,they have to find places to store all of it. If Carlin were toupdate that routine today, he could make the sameobservation about computer information. It seems thateveryone with a computer spends a lot of time acquiringdata and then trying to find a way to store it.

381


3/4

A typical cloud storage system architecture includes amaster control server and several storage servers,as shownin Fig 3.

For some computer owners, finding enough storagespace to hold all the data they've acquired is a realchallenge. Some people invest in larger hard drives.Others prefer external storage devices like thumb drives

or compact discs. Desperate computer owners mightdelete entire folders worth of old files in order to makespace for new information. But some are choosing to relyon a growing trend: cloud storage.

Figure 3. A typical Cloud Storage system architecture

While cloud storage sounds like it has something to dowith weather fronts and storm systems, it really refers to

saving data to an off-site storage system maintained by athird party. Instead of storing information to yourcomputer's hard drive or other local storage device, yousave it to a remote database. The Internet provides theconnection between your computer and the database.

On the surface, cloud storage has several advantagesover traditional data storage. For example, if you storeyour data on a cloud storage system, you'll be able to getto that data from any location that has Internet access.You wouldn't need to carry around a physical storagedevice or use the same computer to save and retrieve yourinformation. With the right storage system, you couldeven allow other people to access the data, turning apersonal project into a collaborative effort.

So cloud storage is convenient and offers more

flexibility, but how does it work? Find out in the nextsection.

III. CLOUD STORAGE REFERENCE MODELThe appeal of cloud storage is due to some of the same

attributes that define other cloud services: pay as you go,the illusion of infinite capacity (elasticity), and thesimplicity of use/management. It is therefore importantthat any interface for cloud storage support theseattributes, while allowing for a multitude of business casesand offerings, long into the future.

The model created and published by the StorageNetworking Industry Association ,shows multiple typesof cloud data storage interfaces able to support bothlegacy and new applications. All of the interfaces allowstorage to be provided on demand, drawn from a pool ofresources. The capacity is drawn from a pool of storagecapacity provided by storage services. The data services

are applied to individual data elements as determined bythe data system metadata. Metadata specifies the datarequirements on the basis of individual data elements oron groups of data elements (containers).

As shown in Fig 4, the SNIA Cloud Data ManagementInterface (CDMI) is the functional interface thatapplications will use to create, retrieve, update and deletedata elements from the cloud. As part of this interface theclient will be able to discover the capabilities of the cloudstorage offering and use this interface to managecontainers and the data that is placed in them. In addition,metadata can be set on containers and their contained dataelements through this interface.

Figure 4.Cloud Storage reference model

It is expected that the interface will be able to beimplemented by the majority of existing cloud storageofferings today. This can be done with an adapter to theirexisting proprietary interface, or by implementing theinterface directly. In addition, existing client libraries suchas XAM can be adapted to this interface as show in Figure4.

This interface is also used by administrative andmanagement applications to manage containers, accounts,

security access and monitoring/billing information, evenfor storage that is accessible by other protocols. Thecapabilities of the underlying storage and data services areexposed so that clients can understand the offering.

Conformant cloud offerings may offer a subset ofeither interface as long as they expose the limitations inthe capabilities part of the interface.

IV. ADVANTAGES AND CHALLENGESWith everything, as has been said before, the devil is

in the details. Certainly, there are more examples of the

382


4/4

growing popularity of cloud storage and valid businessreasons for its popularity. Here are five key benefits ofusing cloud storage and of applications that takeadvantage of storage in the cloud.

Ease of management: The maintenance of thesoftware, hardware and general infrastructure tosupport storage is drastically simplified by an

application in the cloud. Applications that takeadvantage of storage in the cloud are often fareasier to set up and maintain than deploying anequivalent service on premise. At the customersite, often all that is needed to manage yourstorage implementation is a simple web browserleaving the headaches to the service provider.

Cost effectiveness:For total cost of ownership,cloud storage is a clear winner. Elimination of thecostly systems and the people required to maintainthem typically provides organizations withsignificant cost savings that more than offset thefees for cloud storage. The costs of being able toprovide high levels of availability and thescalability an organization needs are also

unmatched. The economies of scale achieved bydata centers simply cant be matched by all but thevery largest of organizations.

Lower impact outages and upgrades:Typicallycloud computing provides cost effectiveredundancies in storage hardware. This translatesinto uninterrupted service during a planned orunplanned outage. This is also true for hardwareupgrades which for the end user will no longer bevisible.

Disaster preparedness: Off site storage isnt new.Keeping important data backed up off site hasbeen the foundation of disaster recovery since theinception of the tape drive. Cloud storage servicesnot only keep your data off premise, but they also

make their living at ensuring that they haveredundancy and systems in place for disasterrecovery.

Figure 5. Survey on concerns with cloud storge services

Simplified planning: Cloud storage solutions freethe IT manager from detailed capacity planning.Cloud-based solutions are flexible and providestorage as needed. This eliminates the need to overprovision for storage that may be needed to meet

However, with every type of cloud storage, there arechallenges in the implementation (i.e. the devil is in the

details). These challenges include: Security (always an issue and not necessarily a

cloud storage specific issue)

Data integrity (making sure the stored data iscorrect)

Power (since you have copies you will have extrastorage which adds power)

Replication time and costs (how fast can youreplicate data since this can be important to dataresiliency)

Cost (how much extra money do you have to payto buy the extra storage for copies)

ReliabilityAs show in Fig 5, the problem include security,control,

performance,support,vendor lock-in, are concerned by

users with cloud storge services.

V. CONCLUSIONS AND FUTURE WORKCloud Storage with a great deal of promise, arent

designed to be high performing file systems but ratherextremely scalable, easy to manage storage systems. Theyuse a different approach to data resiliency, Redundantarray of inexpensive nodes, coupled with object based orobject-like file systems and data replication (multiplecopies of the data), to create a very scalable storagesystem.

This article gives a quick introduction to cloud storage.It covers the key technologies in Cloud Computing andCloud Storage,several different types of clouds services,and describes the advantages and challenges of Cloud

Storage after the introduction of the Cloud Storagereference model.

ACKNOWLEDGMENT

Funding for this research was provided in part by theScientific Research Program of Zhejiang EducationalDepartment under Grant No.20071371. We like to thankanonymous reviewers for their valuable comments.

REFERENCES

[1] Luis M.Vaquero,Luis Rodero-Merino, Juan Caceres,Maik Lindner.A Break in the Clouds: Toward a Cloud Definition. ACMSIGCOMM Computer Communication Review,2009,39(1):50-55.

[2] Wikipedia. Cloud computing .Retrieved fromhttp://en.wikipedia.org/wiki/Cloud_computing,2010.

[3] Wu Jiyi,Ping Lingdi,Pan Xuezeng.Cloud Computing: Concept andPlatform,Telecommunications Science,12:23-30,2009.

[4] Jonathan Strickland.How Cloud Storage Works[OL],http://communication.howstuffworks.com/cloud-storage.htm,2010.

[5] Storage Networking Industry Association.Cloud Storage ReferenceModel,Jun.2009.

[6] Jeffrey B.Layton.Cloud Storage Concepts and Challenges[OL],http://www.linux-mag.com/cache/7617/1.html,2010.

[7] Storage Networking Industry Association.Cloud Storage for CloudComputing,Jun.2009.

383

cloud storage as the infrastructure of cloud computing

Documents